Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
No. 1 Proc. Japan Acad. , 77, Ser. B (2001) 7
Centimeter-level positioning on the seafloor
By Akira AsADA*)'t) and Tetsuichiro YABUKI**)
(Communicated by Yoshibumi TOMODA, M.J.A., Jan. 12, 2001)
Abstract: This paper reports the activity in the development of the observation system of seafloor
crustal deformation for understanding mechanism of interplate earthquakes at subduction plate boundaries
around Japan. Currently, the method for the seafloor geodesy is under development in order to overcome
several difficulties. The position of the sea-bottom transponders at the vicinity of the Nankai trough and the
Japan Trench is observed using a GPS/Acoustic measurement system. Early results from processed data
show a scatter of 4 cm standard deviation for repeated horizontal positioning from drifting survey vessel.
This paper describes two techniques developed to realize precise acoustic ranging, automatically detecting
and correcting the Doppler-shift of acoustic propagation signal due to ship's drifting and attitude motion, and
theoretically forming a pulse of one wavelength in cross correlation between transmission and reception of
long-coded signals.
Key words: Subduction zone; crustal deformation; GPS; acoustic ranging.
Introduction. The space geodetic techniques
have already been used in the monitoring of crustal
deformations on land not only with high spatial and tem-
poral resolutions but also with centimeter-level accuracy. In Japan, the continuous observation with the Global
Positioning System (GPS) Earth Observation Network
(GEONET) has been producing amazing results. These results have contributed significantly to the under-
standing of the complex crustal deformation processes.'
The remote sensing from flying vehicles with SAR
(Synthetic Aperture Radar) Interferometry technique also achieved great success with its high performance.2~
However, these space techniques are only available on
land. Currently, about 70% of the solid earth surface is
under the veil of sea water. Around Japan, many great earthquakes have
occurred repeatedly beneath the deep ocean floor, at the
seismic coupling zone between the descending oceanic
crust and overhanging continental crust along the Kuril Trerlcli, the Japan Trench, the Nankai trough, etc.
When little or no information is known as to what will
happen at the ocean bottom crust, an unexpected
attack of a tsunami generated by a great earthquake can
*) Institute of Industrial Science, The University of Tokyo, 4-6-
4, Komaba, Meguro-ku, Tokyo 153-8505, Japan. **) Hydrographic Department
, Japan Coast Guard, 5-3-1, Tsukiji, Chuo-ku, Tokyo 104-0045, Japan.
S Correspondence to: A . Asada.
be one of the most fearful natural phenomena for the
people who live in the coastal zones.
It is hoped that the observation of sea-bottom
crustal deformations will promote understanding of the
mechanisms of great interplate earthquakes. With the
investigation of seafloor eri_Istal_ dynamics based on the
observed data, more reliable evaluation of the earth-
quake occurrence provability at the plate boundary can be achieved.
South off the shoreline of southwestern Japan,
there exists a plate boundary along the Suruga-Nankai
trough where the Philippine Sea plate is descending
beneath the continental plate. At this subduction plate
boundary, severe earthquakes have occurred repeatedly
with the recurrence interval of about a few hundred
years.3~ The plate convergence rate is estimated at 3-5 cm/year4~; therefore, small relative motion is expected to
be detected from the observation between the coast and
ocean bottom on the landside of the Nankai trough plate
boundary.
In contrast to the Nankai trough, along the Japan
Trench plate boundary, the Pacific Oceanic plate is
descending beneath the continental plate. The conver-
gence rate is estimated about 8 cm/year.5~ Although the severe earthquakes have occurred repeatedly at the
vicinity of the Japan Trench, the distributions of rup-
tured area and the recurrence interval seem less sys-
tematic than those at the Nankai trough. The observation
8 A. ASADA and T. YABUKI [Vol. 77(B),
of sea-bottom crustal deformation is expected to reveal
the inhomogeneity of plate coupling intensity since
most of the seismogenic zone for interplate earth-
quakes along the Japan Trench is at sub sea-bottom.
GPS/Acoustic measurement system for sea-
bottom geodesy. Approach for sea-bottom geodesy is
promoted at several institutions .6)-'0) It is essential to
develop sea-surface precise positioning with GPS and
precise acoustic ranging between sea-surface platform and sea-bottom acoustic transponders (Fig. 1). This is
analogous to GPS geodesy on land. In this GPS/Acoustic method, the precise positioning for the platform and the
buoy on the sea-surface is realized by kinematic GPS
technique which uses phase measurement of carrier
Fig. 1. Schematic image of observation of sea bottom crustal deformation with the
vicinity of convergence plate boundary.
GPS/Acoustic measurement system at the
Fig. 2. Configuration of the GPS/Acoustic measurement system for precise acoustic rang
transducer and sea-bottom transponders.
ing between the ship's
No. 1] Centimeter-level positioning on the seafloor 9
waves as well as code based pseudo range measure- 11)12)
The advantages of GPS/Acoustic positioning, for
the observation at deep sea-bottom around subduction
zones are as follows: (1) The sea-bottom precise
acoustic transponder is simple in design due to its simple function. (2) Neither underwater cable connection nor
acoustic data telemeter is necessary (i.e. these may be
very expensive and/or unreliable) . (3) It is easier to establish a network of geodetic stations on the seafloor.
With this method for realizing sub-meter geodesy
for ocean bottom, four requirements for data collec-
tion/processing are necessary. First, precise positioning
of GPS antenna is required. Secondly, the attitude mea-surement of sea-surface platform enables one to estimate
the position of acoustic transducer from the GPS
antenna position. The third requirement is the record of waveform of acoustic signal that experienced the round
trip between sea-surface and sea-bottom transducers.
The goal is to bring about new techniques superior to the
present range measuring techniques. The fourth requirement is the measurements of sound speed
structure of water which mainly depends on the depth
but also depends on horizontal differences with time to
certain extent. This system is complicated. To establish the system,
we need not only the strict techniques for each mea-
surement, but also the technique to link each of the
instruments systematically. Additionally, the data of from the four requirements need to be processed after
careful consideration of the natures of the error
sources inherent in each type of data so as to obtain highly accurate data. In this report the efforts relating to
the processing of acoustic ranging data will be presented. Attitude measurements of the sea-surface
transducer. The movement of the sea-surface trans-
ducer should be measured within a few centimeters for sufficient accuracy. To accomplish this measurement, a
transducer, a dynamic motion sensor (TSS 335B), a kine-
matic GPS antenna (Trimble 4000SSi) and a GPS head-ing meter (SERCEL NR230 Mk II) are attached to a long
solid pole 8 meters long with a rotational joint (Fig. 2).
The rotational joint is fixed during measurements. It is
useful for raising the transducer while cruising and the GPS signals can be received for a long period of time
without interruption. The GPS heading meter is useful in
measuring the heading of the pitch axis of the dynamic motion sensor.
An attitude log unit was specially designed and
manufactured in order to recode the information from
the dynamic motion sensor, the GPS heading meter and
the transmission-time measurement unit based on GPS
time. It has four RS-232C ports that can be used to attach the CPU time tag to an input block with 1 ms
accuracy. Through the attitude log unit, the motion information can be adjusted to GPS time. Note that com-
munication delay and sensor delay are both taken into
account. For instance, time difference is measured with
1 ms accuracy between a transmission pulse into the
transmission-time measurement unit and a RS-232C output from one.
Precise acoustic ranging technique. The re-
transmitting technique of a received signal by a
transponder is suitable for the realization of precise acoustic ranging.13~ However, to avoid the blunders in
measurement caused by the overlapping of close multi-
path signals, signal deformation due to Doppler-shift,
propagation and background noises, the mirror ability in the sea-bottom station is significantly improved. Mirror
transponder MT3000 manufactured in cooperation with LinkQuest can re-transmit the coded signal with a
204ms span which is the same as the wave it received.
This is analogous to the mirror in optical ranging. With this type of transponder, undetectable blunders can be
removed when the sea-bottom instrument automatically
deals with the signal.
This system consists of a sequence of biphase modulation signal at 10 kHz frequency is used for the
round trip acoustic ranging. The wavelength is about 15
cm. The sequence code of the signal is 255 bits of maxi-
mum length sequence code (M-sequence code). Each
M-sequence chip is modulated with eight carrier waves to provide sufficient stability which results in a signal
length of 204 ms. On sea-surface platform, the received
signal from the sea-bottom transponder, together with a
1 pps (pulse per second) signal from the GPS receiver and the transmission pulse, is recorded on a wave
recorder at a sampling rate of 200 kHz. In principle, the
cross correlation between the replica of the transmitted waveform and the received waveform is computed for
the pulse modulation.
In order to improve tl-te reliability of acoustic rang-ing, two new techniques in the pulse modulation algo-
rithm are developed. One technique uses the most suit-
able transformation of the transmission replica to theoretically form a pulse of one wavelength (Fig. 3).
This is in comparison with the standard replica forming
short bursts of fifteen wavelengths in cross correlation.
This improvement plays a part in fixing the major peak-
point of short bursts and is useful for removing multi-
10 A. AsAVA and T. YABUKI [Vol. 77(B),
path errors. Another new technique deals with Doppler-shifted, ranging signal when the sea-surface transducer is
heaving and drifting. Since the result of the cross corre-lation computation is very sensitive to Doppler-shifts, the
degree of Doppler-shift must be taken into account. In
order to correctly detect a primary pulse and to obtain
accurate short bursts, a lot of replicas are prepared
ahead of time that are equivalent to a wide range of hypothetical shifts. This allows one to search for the best
profit replica at every reception (Fig. 4). Positioning result of sea-bottom transpon-
ders. In February 2000, we have established a sea-bot-tom reference system on the Kumano trough, at 2,000 m
depth, located on the landside of the Nankai trough sub-duction plate boundary, 70 km to the south of the Sima
peninsula, central Japan (Fig. 5). This reference system consists of three mirror type acoustic transponders
(MT-3000) which have been arranged in the form of a right-angled isosceles triangle on the flat seafloor, with
two sides of about 1,400 m length. The reference stations for long base line kinematic GPS positioning using a soft-
ware made by Dr. Colombo are located at two points on
land, Shimosato and Daiozaki; the distances are 70 and
100 km from the sea-bottom geodetic stations. Following the system establishment, three observations
on the sea-bottom reference system have been continu-
ally carried out in February, May and August 2000.
Since the total sea-surface system had been completed, all the measurements were successful on the second
observation. It brought about advanced techniques for
positioning the sea-bottom transponders. Furthermore, two sea-bottom reference systems have been estab-
lished on the landside of the Japan Trench, with the depth of 2,300 m and 2,400 m, in July 2000.
We have analyzed the May observation data of the
Kumano reference system with the advanced signal-pro-
cessing techniques. The measurements of acoustic ranging was carried out while the ship was drifting
(Fig. b). Roundtrip acoustic signals between the sea-sur-face and the sea-bottom units, attitude measurements and the long base line kinematic GPS positioning were
completely adjusted in time and space. In spite of the
severe sea condition, such as ship's vertical motion of +1-2 m (Fig. 7), the scatter of horizontal positioning resid-
uals of three sea-bottom units have achieved about 4 cm
in standard deviation each (Fig. 8). To accomplish this
task, the geodetic position of transducers on sea-bottom
Fig. 3. Demonstration of the improved pulse modulation. Top: The
head part of ranging signal. Middle: The result of standard pulse
modulation using the signals's auto-correlation. Bottom: The
result for obtaining narrow pulses following an improved
method for computing cross correlation between the transmitted
signal and modified replica of the signal.
Fig. 4. While the Doppler-shifted, ranging signal at the sea-surface
transducer is heavung and druting, a primary pulse call be
detected and accurate short bursts can be obtained by preparing
a lot of replicas ahead of time. This is equivalent to having a wide
range of hypothetical shift which allows one to search for the best
proft replica at every reception.
Fig. 5. Positions of the sea bottom geodetic reference system estab- lished with three transponders in February 2000 (triangle), and
GPS reference stations (circles).
No. 1] Centimeter-level positioning on the seafloor 11
was estimated by using the scheme of acoustic ray trac-
ing in the 1 meter mufti-layered sea water structure
based on the WGS84 ellipsoidal earth model. The sound
speed profile was observed with a CTD profiler and a
sound speed profiler.
Concluding remarks. The complexity of sound
speed structure is another important error source for the
ranging between the sea-surface and the sea-bottom.
One method to overcome this error is to obtain various
range measurements with three or more stations on the
seafloor from the sea-surface station drifting through the
center vicinity. With these accurate range measure-
ments and CTD/XBT observation at the center vicinity,
the temporal changes of underwater sound speed profile
were estimated with sufficient accuracy for acoustic posi-
tioning in the center vicinity; although, the horizontal
inhomogeneity may remain a problem to be resolved.
Due to the success of measurements at the
Kumano trough station near the Nankai trough, two sta-tions (two virtual geodetic marks at the center of four
corresponding transponders placed at the square corner)
at the landside of the Japan Trench, with the depth of 2,300 m and 2,400 m, in July 2000 were established. In
succession, several stations are now being deployed
around Miyake volcanic island and the Japan Trench. By repeating the survey with these geodetic stations for sev-
eral years, the sea-bottom crustal deformations should
provide valuable quantitative results. Acknowledgments. On-board observations has been carried out in cooperation with the captains and
crews of the survey vessels "Meiyo" and "Kaiyo". We
would like to thank Dr. Oscar L. Colombo for obtaining highly accurate positioning of the surface acoustic station
using his very long baseline kinematic GPS positioning
software. We gratefully acknowledge helpful discus-
sions with Dr. Xiaolong Yu, Dr. Ning Xiao and Mr.
Fig. 6. The transducer trajectory due to ship's drifting during the
acoustic ranging. In May 2000, at the Kumano trough in the vicin-
ity of the Nankai trough. Circles are the position of acoustic
transponders on the sea-bottom.
Fig. 7. Ship's vertical motion measured by the long base line kine-
matic GPS positioning during the acoustic ranging L1 in Fig. 6.
Fig. 8. The scatter of horizontal positioning residuals of three sea-bottom stations; (a) St. I, (b) St.2 and (c) St.3 in Fig. 6.
12 A. ASADA and T. YABUKI [Vol. 77(B),
Takuro Nakagawa on the transponder system. This study on the Kumano trough sea-bottom station is car-
ried out under the project of Japan Science and
Technology Agency.
References
1) Miyazaki, S., and Hatanaka, Y. (1997) EOS Trans., American (Jeophys, Union 78, no. 17, p. 5104.
2) Massonnet, D,, Rossi, M., Carmona, C., Adragna, A., Peltzer, G., Feigl, K., and Rbaute, T. (1993) Nature 364,138-142. 3) Ando, M. (1975) Tectonophysics 27,119-140. 4) Seno, T., Stein, S., and Gripp, A. E. (1993) J. Geophys. Res.
98,17941-17948. 5) Seno, T., Sakurai T., and Stein, S. (1996) J. Geophys. Res. 101,
11305-11315. 6) Spiess, F. N., and Hildebrand, J. A. (1995) EOS Trans.,
American Geophys. Union 76, no.45, p. 451, p. 455. 7) Spiess, F. N., Chadwell, C. D., Hildbrand, J. A., Young, L. E., Percell, G. H., Jr., and Dragert, H. (1998) Phys. Earth Planet. Inter. 108, 101-112.
8)
9)
10)
11)
12)
13)
Obana, K. (1998) Development of Seafloor Positioning
System with GPS-acoustic Link for Custal Dynamics
Observation, Dr. Thesis, Kyoto University, Japan, pp. 1-99.
Asada, A., and Yabuki, T. (1999) Proc. Inter. Workshop on
recurrence of great interplate earthquakes and its mecha-
nism, Kochi, Japan, pp. 65-75.
Osada, Y., Fujimoto, H., Koizumi, K., Kanazawa, T., and
Murakami, H. (2000) Proc. UT2000, Tokyo, Japan, pp. 306-
314.
Colombo, 0. L. (1998) In GPS for Geodesy (eds. Teunissen, P.
J. E., and Kleusberg, A.). 2nd ed., Springer, New York, pp.
537-568.
Colombo., 0. L., Evans, A. G., Vigo-Aguiar, M. I., and
Benjamin, J. J. (2000) Proc. ION GPS 2000, Salt Lake City,
U.S.A.
Spiess, F. N., Boegeman, D. E., Pavlicek, F. V., and
Lowenstein, C. D. (1980) Precision transponder and
method of communicating therewith, U.S. Patent 4, 214,
314.